pepy.tech

embedding-condensation

Measure layer-wise token embedding cosine similarity, assessing the severity of embedding condensation. Concept from [ICML 2026] Dispersion loss counteracts embedding condensation and improves generalization in small language models.

embedding-condensation has been downloaded 5,313 times in total on PyPI, including 518 in the last 30 days. The latest version is 2.0, released May 17, 2026.

Version2.0
Downloads
5.31k
LicenseNon-Commercial License — see ../LICENSE.md in the LM-Dispersion repository.
AuthorChen Liu
UpdatedMay 17, 2026
Source →PyPI →

Downloads

Weekly, last 90d.
Includes CI traffic.

VersionsTotal2.*1.*
Range
View
Granularity
Group by
CI traffic
Stack: OffCI: Included3 / 5 series
Selected total9.73k
5.3kAll-time
518Last 30 days
2Last 24 h
<0.01/sPer second

Version distribution

Share of downloads by released version. Computed over the last quarter.

  • 01

    2.0

    2.0k downloads
    38.6%
  • 02

    1.1

    1.4k downloads
    25.9%
  • 03

    1.0.0

    992 downloads
    18.7%
  • 04

    0.1.0

    897 downloads
    16.9%

Guess the next day

Thirteen recent days of embedding-condensation downloads. Drag the green handle on the right to guess where day fourteen lands.

TRUTH26MONTUETHUFRISATSUNMONTUEWEDTHUFRISATSUN